> As for Celko's argument that it violates SQL theory, actually it supports,
it:
>
> rows are not supposed to have any order
It would seem to me that a timestamp, which I believe Celko was quoted as
mentioning in the first post in this thread, is just as guilty of violating
this "rule" as an identity field is: a timestamp, or even an integer number
generated by the concatenation of year, month, day, minute, etc, is still
incremental in nature. In other words, if I add two records, the second
record is guaranteed to have a greater ID value than the first whether I'm
using an identity or autonumber field or a timestamp.
Just because I use identity fields does not, in and of itself, impose any
special restrictions or relationships between my data. For example, a record
with an identity field value of 5 is not necessarily preceded by a record
with the value of 4 and followed by a record with the value of 6. By nature
of the identity field the data may have been that way at entry time, but any
one of the records could have been deleted at some time.
I understand Ceklo's point that an incremental identity field may go against
"pure" SQL theory by granting a certain ordering to the data. However, as
long as I don't make the assumption that this ordering will always hold
true, and nothing I do imposes any dependencies on how one identity value
relates to another, the incremental nature can be treated as a coincidental
occurrence. In other words, I may know in the back of my mind that each new
record will have an identity value one greater than the preceding record.
If I ignore that fact, and treat each identity value as nothing more than a
unique handle to the record its assigned to, then I'm no worse off than if I
was using another sort of value for the key. Also, like "Papa Dave" pointed
out, I have the added benefit of NOT having to worry about changes to
natural PKs or PK values duplicated in a foreign table.
To summarize, I too think that identity fields offer enough benefits to
outweigh any "damage" they do to the conceptual model. On a final note, I've
really enjoyed this thread so far: both camps have presented logical
arguments in a very civil discussion (not like most "religious war" style
discussions). My thanks to everyone who's participated: this may be semi OT,
but I think this type of thread is part of what makes this list so great.
Regards,
Seth Petry-Johnson
Argo Enterprise and Associates
------------------------------------------------------------------------------
Archives: http://www.mail-archive.com/cf-talk@houseoffusion.com/
To Unsubscribe visit
http://www.houseoffusion.com/index.cfm?sidebar=lists&body=lists/cf_talk or send a
message to [EMAIL PROTECTED] with 'unsubscribe' in the body.
